linux-kernel - Re: [PATCH v2 16/18] arm64: crypto: disable LTO for aes-ce-cipher.c

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20171120212941.051d11c4@delmar.localdomain>
Date:   Mon, 20 Nov 2017 21:29:41 +0000
From:   Alex Matveev <alxmtvv@...il.com>
To:     Mark Rutland <mark.rutland@....com>
Cc:     Sami Tolvanen <samitolvanen@...gle.com>,
        Andi Kleen <ak@...ux.intel.com>,
        Ard Biesheuvel <ard.biesheuvel@...aro.org>,
        Greg Hackmann <ghackmann@...gle.com>,
        Kees Cook <keescook@...omium.org>,
        linux-arm-kernel@...ts.infradead.org, linux-kbuild@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Masahiro Yamada <yamada.masahiro@...ionext.com>,
        Maxim Kuvyrkov <maxim.kuvyrkov@...aro.org>,
        Michal Marek <michal.lkml@...kovi.net>,
        Nick Desaulniers <ndesaulniers@...gle.com>,
        Yury Norov <ynorov@...iumnetworks.com>,
        Matthias Kaehlcke <mka@...omium.org>
Subject: Re: [PATCH v2 16/18] arm64: crypto: disable LTO for aes-ce-cipher.c

On Mon, 20 Nov 2017 15:20:14 +0000
Mark Rutland <mark.rutland@....com> wrote:

> On Wed, Nov 15, 2017 at 01:34:26PM -0800, Sami Tolvanen wrote:
> > CONFIG_LTO_CLANG requires the use of clang's integrated assembler,
> > which doesn't understand the inline assembly in aes-ce-cipher.c.
> > Disable LTO for the file to work around the issue.  
> 
> Could you elaborate on what the integrated asembler doesn't like?
> 
> It's not entirely clear at a glance, as the asm in that file doesn't
> seem to do anything that obscure.
> 
> Is it a bug?

Turns out, integrated assembler doesn't like this instruction in
aes_sub():
	umov	%w[out], v0.4s[0]

Specifically, it barks at "v0.4s[0]" part. And the way I read the spec,
it's quite correct in not accepting this argument. From UMOV
description:

UMOV <Wd>, <Vn>.<Ts>[<index>]

...

<Ts> For the 32-bit variant: is an element size specifier, encoded in
the "imm5" field. It can have the following values:
	B	when imm5 = xxxx1
	H	when imm5 = xxx10
	S	when imm5 = xx100


With "v0.s[0]" it builds fine.

Ard, since this is your code, can you comment? Feels like a typo.

Regards,
Alex